December 18, 2012 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
December 18, 2012 [00:06:18] *** kfr- has quit IRC[00:06:32] *** wolfeidau has quit IRC[00:06:44] *** e^ipi has quit IRC[00:07:15] *** e^ipi has joined #smartos[00:07:27] *** wolfeidau has joined #smartos[00:08:53] *** des2 has quit IRC[00:13:32] *** xer0x has quit IRC[00:16:41] *** des2 has joined #smartos[00:21:37] *** rancor has quit IRC[00:29:12] *** wramthun2 has joined #smartos[00:31:11] <jbergstroem> konobi: thanks a lot[00:31:56] *** wramthun3 has quit IRC[00:42:06] *** xer0x has joined #smartos[00:52:20] *** xer0x has quit IRC[00:54:38] *** neophenix has quit IRC[00:57:28] *** wramthun2 has quit IRC[00:57:52] *** wramthun2 has joined #smartos[00:58:16] *** xer0x has joined #smartos[01:03:54] *** tonyarkles has quit IRC[01:10:46] *** gkyildirim has quit IRC[01:23:58] <konobi> jbergstroem: that work?[01:28:44] *** tonyarkles has joined #smartos[01:29:14] *** jim80net has quit IRC[01:34:32] *** robinsmidsrod has quit IRC[01:40:55] *** robinsmidsrod has joined #smartos[01:43:39] *** ira has quit IRC[01:43:46] *** jim80net has joined #smartos[01:52:45] *** wdent has quit IRC[02:02:12] *** iyp has joined #smartos[02:03:03] *** xer0x has quit IRC[02:05:39] *** jim80net has quit IRC[02:06:55] *** wdent has joined #smartos[02:14:07] <jbergstroem> konobi: unfortuantely no :([02:19:25] <konobi> yeah, not sure what you can do if it's not a clean poweroff[02:29:32] *** iyp has quit IRC[02:30:52] *** xer0x has joined #smartos[02:31:27] <jbergstroem> me neither[02:32:34] <konobi> mkay... intel nic installed... blog time[02:39:30] *** bixu has joined #smartos[02:39:58] <bixu> How can I identify where 'sdN' is mounted from within my zone?[02:40:13] <bixu> (Can't see the global as I'm a JPC customer...)[02:40:42] *** xer0x has quit IRC[02:40:50] <konobi> bixu: smartos zone?[02:40:57] <bixu> Yes.[02:41:00] <bixu> Not KVM.[02:41:02] <konobi> you can't[02:41:12] <bixu> Uhm.[02:41:25] <bixu> I want to know what 'sd' device is '/' - I can't know that?[02:41:46] <konobi> nope, because it's not really a device[02:41:47] <rmustacc> There is no correlation between them.[02:41:56] <wesolows> normally there's no 1-1 mapping of disks onto filesystems[02:42:03] <bixu> I see.[02:42:44] <bixu> So I should just use -z to get cleaner output?[02:43:13] <rmustacc> If you're in a zone, looking at the disks isn't what you want.[02:43:17] <wesolows> from what utility?[02:43:28] <rmustacc> If you want to see that you have filesystem induced latency you should use vfsstat.[02:43:38] <bixu> rmustacc: Ok.[02:44:04] <bixu> I had been told by support once to use 'iostat -x N' and look at the 'b' column.[02:44:07] <rmustacc> What is the actual piece of information you want or problem you want to solve?[02:44:15] <rmustacc> Um.[02:44:44] <bixu> A few things. Mostly just trying to be sure that I'm not hitting an I/O wall as I debug what I think is a code issue.[02:44:59] <rmustacc> Then use vfsstat.[02:45:04] <bixu> Okay.[02:45:09] <rmustacc> That'll give your fs latency.[02:45:17] <bixu> Thanks - and that makes sense.[02:45:27] <rmustacc> Good.[02:46:54] <bixu> read_t and write_t are milliseconds?[02:47:01] *** joe_d_ has quit IRC[02:47:04] *** joe_d has joined #smartos[02:47:13] <bixu> (in vfsstat output)[02:48:24] <rmustacc> I don't recall, sorry.[02:48:37] <bixu> No worries. Digging now.[02:48:58] <bixu> Looks like ziostsat reports much the same data?[02:49:48] <rmustacc> It looks like it covers some of it, yes.[02:50:37] <bixu> I've found that sometimes it helps web developers if you don't give them a tool that edits out some of the more nitty-gritty data. ziostat seems to fit the bill :)[02:52:05] <konobi> which repo is v8.so in?[02:52:07] <rmustacc> You can also use cloud analytics to get a graph of that.[02:52:13] <rmustacc> konobi: it's in ON.[02:52:19] <konobi> ta[02:52:20] <bixu> Ok - according to Bill's blog post on dtrace.org, read_t and write_t give values in ms (milliseconds).[02:57:43] *** tonyarkles has quit IRC[03:02:04] <konobi> rmustacc: how's best to file an issue on illumos-joyent?[03:02:17] <konobi> just upstream illumos-gate?[03:02:51] <wesolows> unless it's specific to illumos-joyent, yes[03:03:03] <konobi> v8.so?[03:03:58] <konobi> wesolows: just an idea to have a 64-bit version of it in /usr/lib/mdb/proc/amd64/[03:04:03] <wesolows> pretty sure that's ours... most non-customers seem to file such bugs against smartos-live on github. it's definitely not right but there's no obviously better option[03:04:44] <wesolows> indeed. not sure why there isn't but I can imagine a few reasons. or it may just be an oversight.[03:05:02] <konobi> probably because the system node is 32bit[03:05:45] <konobi> k... i'll file an issue on smartos-live[03:07:13] <konobi> done[03:08:46] *** khushildep has quit IRC[03:11:45] <konobi> wesolows: along with another one which Josh'll look quizzically at[03:12:07] <konobi> though i get to blame robg for that[03:18:26] *** iyp has joined #smartos[03:26:46] *** iyp has quit IRC[03:31:53] *** dap has quit IRC[03:37:05] <konobi> rmustacc: about?[03:41:44] <konobi> richlowe: actually you may know too[04:25:37] *** relling has quit IRC[04:30:06] <jbergstroem> konobi: tried all kinds of things now; just can't get the stop to return in a timely fashion (obsd 5.2). giving up.[04:33:34] *** darjeeling has quit IRC[04:44:34] *** xmerlin has quit IRC[04:54:54] <chris---> anyone trunking a bunch of vlans to their hosts and giving those out to vms that can answer a few questions[04:55:08] *** xforest has quit IRC[04:56:22] <konobi> chris---: hhhmmm?[04:57:23] <arekinath> chris---: I'm trunking vlans onto hosts... only about 6 vlans, but I'm using them[04:57:28] <chris---> so we have ways we do this in solaris and esx, and i'm trying to make that work in smartos[04:57:44] <chris---> i've been able to make vms which look right, but dont work[04:57:47] <konobi> chris---: add a vlan_id to your vm json[04:57:52] <arekinath> chris---: normally what you do in smartos is put the trunk onto the "external" nic, and use vlan_id[04:57:59] <arekinath> chris---: don't try to separate them out using vlans in dladm by hand[04:58:20] <chris---> trunk_nic=68:5:ca:b:63:69[04:58:21] *** wolfeida_ has joined #smartos[04:58:21] <konobi> chris---: read my blog post?[04:58:31] <chris---> i may have... :) point me at it[04:58:35] *** wolfeidau has quit IRC[04:58:43] <chris---> we do this in solaris 10 and 11, and it looks different[04:58:44] <konobi> cloudtone.wordpress.com[04:59:32] <arekinath> chris---: using vlan_id "just works"[04:59:36] <arekinath> chris---: no need to configure anything esle[04:59:38] <arekinath> *else[04:59:40] <chris---> i've been able to make nics in the global that work[05:00:12] <konobi> chris---: is "trunk" appearing in `sysinfo` ?[05:00:14] <arekinath> so your nic should look something like "nics": [{"nic_tag":"trunk", "vlan_id": 371, "ip": "...", ...}][05:00:28] <arekinath> if you've set trunk_nic in your usbkey config like that[05:00:32] <arekinath> and rebooted afterwards[05:03:24] <chris---> ok so if i dont make a ip/netmask/vlan_id in the usbkey/conf it doesnt appear to configure it[05:03:26] <chris---> sorry kids[05:03:49] <konobi> chris---: does it appear in `sysinfo` ?[05:03:53] <chris---> so if i do that in the global then the nic comes up etc and looks like its working[05:04:02] <chris---> yes[05:04:20] <arekinath> it should work without needing to give the global zone a vnic on it...[05:04:25] <konobi> so you have a NIC that has NIC Names: [trunk] ?[05:04:31] <konobi> arekinath: shush[05:04:38] <arekinath> konobi: sorry[05:04:40] <chris---> "e1000g1": {"MAC Address": "68:05:ca:0b:63:69", "ip4addr": "", "Link Status": "up", "NIC Names": ["trunk"]}[05:05:32] <konobi> chris---: okay... so when you create a VM... for the NIC... use "nic_tag: trunk, vlan_id: <id>"[05:05:42] <chris---> right, i do that[05:05:50] <konobi> that's it... your VM will now be on that vlan id[05:06:24] <chris---> hrm, sure doesnt work though[05:06:48] <konobi> dladm show-vnic will show how the vnic is configured[05:07:08] <chris---> sec i'll redo what i had[05:07:28] <konobi> if it's not working, though those details are correct, then your switch config is borked[05:08:49] <konobi> arekinath: sorry... just wanted step by step answers ^_^[05:09:19] <chris---> net0 e1000g1 0 12:6d:b7:61:8b:19 fixed 847 23d32faa-0d.....[05:09:24] <chris---> from global[05:09:51] <chris---> looks right then?[05:09:56] <konobi> so your vlan id is 847?[05:09:59] <chris---> yeah[05:10:06] <konobi> looks right to me[05:10:21] <konobi> you're not running within virtualbox or vmware, right?[05:10:23] <chris---> it looks different than what we get in sol11 doing zones which is why i was wondering[05:10:28] <chris---> right[05:10:55] <chris---> maybe the way we are doing it in sol11 is weird[05:11:03] <konobi> then yup... that's correct[05:11:35] <chris---> but we end up with a vnic say vlan847, and when the zone starts up say zone1 there is a vlan847\zone1 vnic created also[05:11:40] <chris---> its weird[05:12:15] <chris---> vl1152 net0 1000 2:8:20:4a:ab:b1 random 1152[05:12:16] <konobi> yeah, it's a little overloaded that way[05:12:16] <chris---> est-oc1/vl1152 net0 1000 2:8:20:4a:ab:b1 random 1152[05:12:28] <chris---> so i was expecting something similar[05:13:24] <chris---> so let me reboot this thing without anything but the trunk_nic, leaving out the stuff to make ints in the global[05:14:21] <chris---> i cant say i like what we ended up with in sol11[05:14:52] <chris---> we have to have a vnic per zone, then the zone makes another as it boots[05:14:54] <chris---> seems odd[05:15:15] <chris---> booting from usb, so it will take a bit[05:16:50] <chris---> i've changed out the onboard hp sas cntroller, added intel nics, added mem, this is gonna work... somehow[05:18:16] <chris---> northern ireland huh[05:18:25] <konobi> mmm?[05:18:37] <chris---> bottom of the blog post[05:18:44] <konobi> ah, yes[05:18:53] <chris---> think i've been to sothern ireland[05:19:20] <chris---> cork, limerick, shannon[05:24:26] <chris---> ok so its back up, show-phys shows the trunked interface as in lala land, sysinfo shows the trunk interface[05:25:49] <konobi> should be working[05:27:00] <chris---> ok so once the vm booted that other nic came up and showed link etc[05:27:12] <chris---> vnics created[05:28:02] <chris---> doesnt work though[05:28:31] <chris---> but it sonds like i have what i should have nic/vnic wise[05:28:46] <konobi> time to check on the switch config =0)[05:28:48] <konobi> bbl[05:28:52] <chris---> thanks[05:32:08] *** darjeeling has joined #smartos[05:44:46] *** sachinsharma has joined #smartos[05:55:15] <chris---> ok so it works[05:55:17] <chris---> thanks[05:59:47] <chris---> now i'm curious whether sol11 zones can work the same way, or if there is a fundamental difference[06:07:30] *** linuxprofessor has quit IRC[06:32:55] *** linuxprofessor has joined #smartos[07:14:52] <rmustacc> konobi: I am now[07:16:54] *** wolfeida_ has quit IRC[07:22:38] *** xer0x has joined #smartos[07:22:46] *** xer0x has quit IRC[07:45:16] *** wolfeidau has joined #smartos[07:53:18] *** linuxprofessor has quit IRC[08:25:42] *** ryancnelson has joined #smartos[08:25:45] *** linuxprofessor has joined #smartos[08:28:09] *** alucardX has joined #smartos[08:28:23] <alucardX> morning[08:31:23] *** mamash has joined #smartos[08:37:50] *** kschiess has joined #smartos[08:48:19] *** ryancnelson has quit IRC[08:59:44] *** jamesd_ has quit IRC[09:22:13] *** bens1 has joined #smartos[09:23:10] *** texarcana has quit IRC[09:24:24] *** texarcana has joined #smartos[09:47:00] <jperkin> yawn[09:57:31] <MerlinDMC> morning![10:27:31] *** sachinsharma has quit IRC[10:40:40] *** bayoda has quit IRC[10:45:33] *** bayoda has joined #smartos[10:53:05] *** sachinsharma has joined #smartos[11:28:20] *** elijah-mbp has quit IRC[11:53:36] *** Cpt-Oblivious has quit IRC[11:57:18] *** russiane39 has quit IRC[11:57:35] *** russiane39 has joined #smartos[12:00:38] *** darjeeli_ has joined #smartos[12:04:30] *** darjeeling has quit IRC[12:15:25] *** darjeeli_ has quit IRC[12:15:53] *** darjeeling has joined #smartos[12:20:31] *** darjeeling has quit IRC[12:22:42] *** bayoda has left #smartos[12:46:29] *** mamash has left #smartos[12:58:06] *** kamilr has joined #smartos[12:58:26] <kamilr> Hi there[12:58:46] <kamilr> Can i create image with two disks ?[12:59:26] *** jamesd has joined #smartos[13:07:23] *** darjeeling has joined #smartos[13:18:20] *** alucardX has quit IRC[13:18:25] *** sachinsharma has quit IRC[13:18:40] *** alucardX has joined #smartos[13:19:25] *** wdent has quit IRC[13:27:12] *** sjorge has quit IRC[13:32:23] *** sjorge has joined #smartos[13:32:24] *** sjorge has joined #smartos[13:40:23] *** ira has joined #smartos[13:47:20] *** rc10 has joined #smartos[13:48:20] <rc10> hi, For zfs poo, I have been using mirrored zil but its not getting used. I have huge writes everyday[13:48:55] <rc10> i read up that Zil works only for synchronous writes rather normal writes[13:49:18] <rc10> any pointers would be helpful[14:03:03] *** sachinsharma has joined #smartos[14:03:22] <yofuh> rc10: so, you have the answer already, whats the question?[14:04:28] <yofuh> how you can force sync writes to make use of the zil? this would likely slow down writing insteat of givng it a boost[14:05:29] <MerlinDMC> rc10, do you have any frequently occuring synced writes?[14:10:20] *** wdent has joined #smartos[14:26:44] <rc10> I dont have sync writes just normal writes[14:26:56] <rc10> how to make use of Zil for normal writes[14:27:43] <linuxprofessor> you can't as far as I know[14:30:37] <linuxprofessor> and there's really no point since the write is cached[14:31:50] <linuxprofessor> if you need fast I/O for async writes, consider using flash as main storage[14:32:35] <rc10> fine[14:32:54] <linuxprofessor> sorry to dissapoint you :-)[14:33:08] <linuxprofessor> +p[14:33:45] <norman> ha that nick :)[14:34:08] <linuxprofessor> I'm a professor and I teach a bunch of Linux classes :p[14:34:59] <linuxprofessor> it sure would be nice to have auto tiering in the sense that data is moved between SSDs and HDDs[14:35:12] <linuxprofessor> don't know if it's somewhere in the pipeline?[14:37:22] *** thrashout has joined #smartos[14:40:59] *** thrashout has quit IRC[14:44:02] <nahamu> linuxprofessor: an L2ARC that doesn't need to warm up on every reboot would come pretty close (though it would duplicate data...)[14:44:24] <MerlinDMC> rc10, normal writes should be cached in ram already[14:44:25] <nahamu> I think I heard slight mutterings about that, though I have no idea if anyone is working on it.[14:45:06] <linuxprofessor> true, the warm up can be kind of irritating[14:46:08] <linuxprofessor> it usually takes about 24 hours to warm up the l2arc[14:46:24] <linuxprofessor> so it's kind of long time without "optimal" performance[14:55:07] *** sachinsharma has quit IRC[14:58:24] *** bens1 has quit IRC[15:00:43] *** jim80net has joined #smartos[15:04:16] *** iyp has joined #smartos[15:10:23] *** KermitTheFragger has joined #smartos[15:20:41] *** elijah-mbp has joined #smartos[15:30:26] *** relling has joined #smartos[15:40:37] *** CarlosC has joined #smartos[15:49:38] *** rc10 has quit IRC[15:53:59] *** eule has joined #smartos[15:55:05] *** kschiess_ has joined #smartos[15:57:12] *** kschiess has quit IRC[15:58:28] *** eule has quit IRC[15:59:28] *** linuxprofessor has quit IRC[16:02:43] *** newlix_ has joined #smartos[16:17:32] *** elijah-m_ has joined #smartos[16:20:29] *** elijah-mbp has quit IRC[16:22:17] *** newlix_ has quit IRC[16:22:25] *** neophenix has joined #smartos[16:32:13] *** xmerlin has joined #smartos[16:37:16] *** neophenix_ has joined #smartos[16:38:38] *** daleg_ has joined #smartos[16:38:39] *** daleg has quit IRC[16:38:40] *** daleg_ is now known as daleg[16:40:07] *** ahaydock has joined #smartos[16:40:57] *** neophenix has quit IRC[16:40:58] *** neophenix_ is now known as neophenix[16:43:48] *** rancor has joined #smartos[16:46:32] *** relling has quit IRC[16:53:57] *** tonyarkles has joined #smartos[16:58:13] *** alucardX has quit IRC[16:58:14] *** relling has joined #smartos[17:29:51] *** kamilr has quit IRC[17:46:48] *** sachinsharma has joined #smartos[17:54:48] *** jelmd has quit IRC[17:58:34] *** khushildep has joined #smartos[17:59:22] *** tonyarkles has quit IRC[18:00:51] *** khushildep has quit IRC[18:04:59] *** tonyarkles has joined #smartos[18:09:12] <iyp> I have a package that will build with just "bmake" but I get this error with "bmake package"[18:09:19] *** montyz0 has joined #smartos[18:09:21] <iyp> https://gist.github.com/84587874309289f8400d[18:10:14] <jperkin> iyp: bmake deinstall[18:11:10] <iyp> jperkin: thanks. then just bmake package again?[18:11:15] <jperkin> yep[18:11:20] <iyp> cooll[18:12:27] <iyp> jperkin: https://gist.github.com/24f3d492feeb0fcb2ec8[18:12:45] <iyp> guess I forgot some requirements? Don't know why php needs to be pulled in though...[18:13:43] <jperkin> iyp: CHECK_INTERPRETER_SKIP+=bin/sqlsample.php[18:13:46] *** dap has joined #smartos[18:14:02] <jperkin> it's a PKG_DEVELOPER=yes check to make sure that installed files use the correct tools[18:14:17] <jperkin> though it'd probably be better to not install that file in the first place, if it's just a sample and not useable by default[18:14:35] <iyp> Oh ok. Yeah, I had to set PKG_DEVELOPER to get the PLIST to print right[18:15:22] <jperkin> I'd do something like this[18:15:26] <jperkin> post-install:[18:15:40] <jperkin> ${RM} -f ${DESTDIR}${PREFIX}/bin/sqlsample.php[18:15:43] <jperkin> then remove it from the plist[18:15:54] <iyp> ok[18:16:10] <iyp> some weird error just came out of the woodwork when I added that check interpreter line[18:16:23] *** kschiess_ has quit IRC[18:16:54] <iyp> https://gist.github.com/ca90cc353a68e350e34e[18:17:33] <jperkin> that looks like it didn't parse your mk.conf[18:17:37] <jperkin> does it still exist? :)[18:18:20] <iyp> oh…wow[18:18:21] <iyp> no[18:18:27] <iyp> where the hell did that go...[18:18:47] <jperkin> did you rm ${PREFIX} somehow? :)[18:19:10] <iyp> boy I hope not. /opt/local is still there[18:19:53] <iyp> ok. got the mk.conf from another zone. Running now[18:20:21] <iyp> so do I do both the post-install section and set CHECK_INTERPRETER_SKIP+=bin/sqlsample.php[18:20:27] <iyp> or is it just one or the other?[18:20:32] <jperkin> one or the other[18:20:49] <iyp> you would go with the post-install?[18:21:05] <jperkin> yeh, it doesn't sound like sqlsample.php provides enough value[18:21:28] <jperkin> and, if it was to be packaged, should probably go in ${PREFIX}/share/examples instead[18:21:55] <iyp> I don't care about it. We don't touch php around here.[18:22:21] <iyp> Guess it might be nice to include for people if I publish this package though[18:22:36] <iyp> Which I intend to. What does that entail?[18:24:15] <jperkin> first get it into pkgsrc-wip for people to test (pkgsrc-wip.sf.net)[18:24:42] <jperkin> then once it's tidied up enough one of us can import it into pkgsrc[18:26:00] <iyp> Awesome[18:26:34] *** mamash has joined #smartos[18:28:08] <iyp> jperkin: new fun on the bmake package[18:28:09] <iyp> https://gist.github.com/579f9d9148b1ba130eb1[18:28:38] <iyp> does that mean I have to pull in perl?[18:28:58] <iyp> I already do in USE_TOOLS[18:29:13] <iyp> https://gist.github.com/5efdc60ed5e455b940e4[18:29:21] <jperkin> no, it just means that the package incorrectly figured out the location of perl - again,if they aren't much use then just rm them[18:29:40] *** khushildep has joined #smartos[18:30:53] <jperkin> iyp: remove the CFLAGS line, --prefix=/opt/local and --with-unixodbc=/opt/local[18:31:31] <iyp> remove both those options?[18:31:36] <jperkin> yeh[18:32:21] <iyp> …mk.conf got deleted again[18:32:32] <iyp> I'm doing something terribly wrong somewher[18:33:33] <iyp> all I've run is 'bmake clean' 'bmake deinstall' and 'bmake package'[18:34:17] <jperkin> can you dump the PLIST somewhere?[18:34:39] <jperkin> also, you shouldn't need to 'deinstall' if you are only doing 'package', only if you do 'install' (which actually installs the binary package to $PREFIX)[18:35:40] *** bens1 has joined #smartos[18:38:13] *** khushildep has quit IRC[18:42:01] *** khushildep has joined #smartos[18:43:30] *** montyz01 has joined #smartos[18:43:30] *** montyz0 has quit IRC[18:45:00] *** khushildep has quit IRC[18:55:41] *** ira has quit IRC[18:56:42] <iyp> jperkin: PLIST -- https://gist.github.com/178141abe0e3e74addcd[18:57:36] <jperkin> iyp: ok, that's the problem, you have etc/mk.conf in there :)[18:57:42] <iyp> hahaha[18:57:46] <jperkin> so every time you do 'bmake uninstall' it's getting rm'd[18:57:56] <iyp> well that makes perfect sense.[18:58:14] <iyp> Wonder how it ended up in there. I just used the output of bmake print-PLIST[18:58:30] <jperkin> you probably edited it in between an install and a print-PLIST[18:58:42] <jperkin> all print-PLIST does is find based upon mtime[19:00:52] *** ryancnelson has joined #smartos[19:01:13] <jperkin> there are a few other changes we can make to that PLIST, perl modules can be automatically handled with the .packlist files, and python should be made version-agnostic, etc -> ${PKG_SYSCONFDIR}, and a few others, but I need to go afk (putting it in wip would probably get some folks interested in cleaning it up further)[19:01:41] <jperkin> pkgtools/pkglint is mostly useful for checking a package[19:06:39] *** tonyarkles has quit IRC[19:15:39] *** linuxprofessor has joined #smartos[19:41:35] *** ryancnelson has quit IRC[19:45:08] *** jim80net has quit IRC[19:47:56] *** KermitTheFragger has quit IRC[19:53:35] *** mamash has left #smartos[19:58:10] *** josephholsten_ has joined #smartos[20:03:45] *** riceo has joined #smartos[20:08:44] *** tonyarkles has joined #smartos[20:11:06] *** ryancnelson has joined #smartos[20:22:35] *** jim80net has joined #smartos[20:25:59] *** khushildep has joined #smartos[20:35:38] *** mamash has joined #smartos[20:37:59] *** xer0x has joined #smartos[20:38:36] *** tonyarkles has quit IRC[20:42:42] *** riceo has quit IRC[20:42:56] *** sachinsharma has quit IRC[20:51:05] *** szaydel has joined #smartos[20:51:07] *** szaydel has quit IRC[20:51:37] *** szaydel has joined #smartos[20:52:09] *** szaydel has joined #smartos[20:52:17] *** CarlosC has quit IRC[20:56:13] *** jelmd has joined #smartos[21:14:26] *** jelmd has quit IRC[21:16:27] *** montyz01 has quit IRC[21:16:38] *** montyz0 has joined #smartos[21:31:01] *** bens1 has quit IRC[21:44:03] *** khushildep has quit IRC[21:44:25] *** khushildep has joined #smartos[21:45:47] *** tonyarkles has joined #smartos[21:48:27] *** jamesd has quit IRC[21:50:29] *** dap has quit IRC[21:50:55] *** dap has joined #smartos[21:51:13] <chris---> so can anyone speak to the differences in how i see networking in sol11 and in smartos re vnic and zones/vms? I'm really curious[21:51:25] <chris---> never see the something/somehting vnic in smartos[21:52:09] <ryancnelson> the "something/something" vnic?[21:52:27] <chris---> est-oc1/vl1152 net0 1000 2:8:20:4a:ab:b1 random 1152[21:52:37] <chris---> show-vnic output in s11[21:52:45] <ryancnelson> oh. we don't do that.[21:52:45] <chris---> for the vnic going to est-oc1 zone[21:52:56] <ryancnelson> smartos zones are their own brand[21:53:12] <ryancnelson> … a feature of that brand is dynamically created vnics[21:53:13] <chris---> is that something fundamentally different between the os or just different plumbing[21:53:22] <ryancnelson> when the zone is off, the vnic doesn't exist[21:53:24] <chris---> well thats dynamically created also[21:53:27] <chris---> same with that[21:53:41] <chris---> its almost the same, just not quite[21:54:00] <chris---> and i wasnt sure if that was something that has diverged or just a different way its done[21:54:32] <chris---> its like everything is off by one[21:54:49] <chris---> show-phys has the same thing for DEVICE and LINK in smartos[21:55:09] <chris---> and in sol11 its net# for LINK there[21:55:19] <chris---> and then it seems like its off by one the rest of the way through[21:56:20] *** newlix_ has joined #smartos[21:56:21] *** newlix has quit IRC[21:56:33] <chris---> i'm knee deep in both environments, been interesting seeing the same thing in a slightly different way side by side[21:57:49] <chris---> maybe more frustrating than interesting :)[21:58:50] <chris---> i like the smartos way[21:59:11] <ryancnelson> that's done because you can't predict what order things will probe in, from boot to boot.[21:59:27] <ryancnelson> on the box i'm looking at, link and device are all the same[21:59:36] <ryancnelson> but next boot, they might not be[21:59:48] <ryancnelson> link is consistent.[21:59:49] <chris---> yeah thats never the case on s11, it always starts abstracting there to net#[22:00:03] *** daleg has quit IRC[22:00:28] *** daleg has joined #smartos[22:01:04] <chris---> creating a zonename/vnic entry would be interesting since the zone names are awfully long[22:01:09] <chris---> in smartos[22:02:40] <chris---> in the wiki, is there a way to like submit proposed changes/additions? if you register you dont get free run do you?[22:02:41] <ryancnelson> not sure i understand what you're trying to do. you don't hand-roll zones in smartos. let the system do that.[22:03:37] <chris---> no but as i do things across the two environments I see how things are different, and i'm just trying to understand the different ways and why[22:03:50] <chris---> i'm not a very good "just use it and dont care how it works" person i guess[22:05:02] <chris---> if we could do things the same way across the two environments i would[22:05:10] <ryancnelson> that's fine. just don't be a "i expect it to work just like solaris 11 does" person[22:05:31] <chris---> no its a curiosity as to why it works how it does and understanding thing[22:05:39] <ryancnelson> no, don't. you can't do things the same way across both environments[22:05:40] *** zr0 has joined #smartos[22:05:46] <ryancnelson> it's not free solaris.[22:05:47] *** zr0 has quit IRC[22:06:25] <ryancnelson> it's a thin in-ram os to run vms and zones on[22:06:42] *** zr0 has joined #smartos[22:07:54] <chris---> not trying to make them the same, like i said, if we can learn from one i'm willing. but like you said it doesnt translate[22:09:37] <chris---> interesting to see the same stuff used differently[22:09:41] <ryancnelson> not trying to be opaque here… but if you're using zoneadm and dladm do to anything but look at the state of things, then you're using the system in a way it's not intended. you're right. it doesn't translate.[22:11:29] <ryancnelson> if vmadm doesn't address the thing you're trying to do with your zone, then you're off the reservation, and other things will probably break later when it assumes the state of the world is as vmadm put it.[22:14:25] *** rcamera has joined #smartos[22:24:00] *** newlix has joined #smartos[22:26:35] <chris---> we do a fair amount of solaris and zones etc, so its just kinda neat to see it used for something like this and understand some of how it works. by no means trying to toss out the rules. this isnt a "how can i make this solaris" question :)[22:27:11] *** newlix_ has quit IRC[22:27:12] <chris---> i see your shields going up :)[22:27:39] <chris---> we like it for what it is, just like peeking behind the curtain[22:28:22] <ryancnelson> *every* *single* *day* there's people in here trying to install smartos to disk, put zones on iscsi, run oracle RAC, and port it to ultrasparc :)[22:28:29] <chris---> yeah i see it :)[22:28:34] <chris---> i was trying to not be that guy[22:28:49] <ryancnelson> you're not. you're speaking in complete sentences[22:28:57] <ryancnelson> s'all good.[22:29:25] *** jelmd has joined #smartos[22:29:32] <chris---> in this case it was really "i wonder if s11 could be more like smartos here"[22:29:54] <chris---> tooling up for both at the same time[22:30:04] <ryancnelson> i believe the link order you see in dladm show-phys is always something like "alphabetical by mac addr", to avoid probe-return races at boot. solaris 11 does more complicated tricks like DID's, etc...[22:30:54] <chris---> yeah it seems like s11 is off by one, one level farther down the road by the time it gets to giving the vnic to the zone[22:31:01] *** xer0x has quit IRC[22:31:31] <chris---> and i also think my problems with this broadcom nic may have been misunderstanding of how it all works in smartos[22:31:41] <ryancnelson> ew. which brodcom nic?[22:31:48] <ryancnelson> bnx or bge?[22:32:10] <chris---> bnx[22:32:34] <rmustacc> Well, we also don't have the best drivers for broadcom nics.[22:32:36] <chris---> with no vms up, it isnt shown as up[22:32:39] <ryancnelson> those are known to be quite quirky. closed driver, so the devs can't fix bugs[22:32:43] <chris---> which i think threw me[22:33:31] <ryancnelson> yeah, that's also a thing that confuses a lot of people… the GZ doesn't have a link into every network, even though you can put vm's on those links[22:33:31] <chris---> put a dual port intel nic in it, see the same thing, but figured out the correct behavior isnt what you would expect[22:33:59] <chris---> so i'm gonna go back to the onboard and see if they work now that i know what working looks like[22:34:05] <chris---> just to see[22:34:42] <chris---> have 7 more of these, extra nics add up[22:35:22] <ryancnelson> … yeah, for figuring out if my cables are connected right, i often do a "ifconfig -a plumb up"[22:35:35] <ryancnelson> just so i can see the link status right from the GZ[22:35:58] <chris---> it wouldnt show me the link up ever for the ther ports[22:36:02] <chris---> still may not be right[22:36:13] <ryancnelson> 'cause, like I said, you can hang a vnic off ixgbe3 and let a zone use it, and still see nothing in "ifconfig" from the GZ[22:36:24] <chris---> but ill try armed with seeing it work finally with this nic and just see[22:36:55] <chris---> i can disable them in the bios, and the onboard raid, so at least i can clean that cruft out[22:37:07] <chris---> put in an lsi card and a dual port intel nic[22:37:32] <chris---> turned it into a fairly nice box[22:38:42] <chris---> almost like sunoracle hardware[22:39:00] <chris---> serial doesnt work as well though[22:40:30] <chris---> have gotten all the way down to getting pkgin to update the database which it cant cause its in 10. land[22:40:34] <chris---> so close[22:41:43] *** wolfeidau has quit IRC[22:46:33] <chris---> and now that pkgin works... i seem to finally have a fully functional death star...[22:47:36] <chris---> time to make freebsd vms booting zfs in a zone on top of zfs and watch the world implode[22:49:29] <ryancnelson> ouch… copy-on-write on copy-on-write will hurt perf[22:50:17] <chris---> probably best to stay one step away from implosion[22:52:29] *** wdent has quit IRC[23:00:43] *** relling has quit IRC[23:03:38] *** montyz0 has quit IRC[23:07:01] *** xer0x has joined #smartos[23:07:24] *** D-Boy has joined #smartos[23:08:23] *** wolfeidau has joined #smartos[23:09:11] *** tonyarkles has quit IRC[23:11:11] *** bens1 has joined #smartos[23:11:12] *** relling has joined #smartos[23:11:12] *** josephholsten_ has quit IRC[23:11:29] *** josephholsten_ has joined #smartos[23:15:01] *** bens1 has quit IRC[23:15:49] *** D-Boy has left #smartos[23:16:03] *** kfr- has joined #smartos[23:16:39] <xmerlin> hi to all, I've a strange imgadm error http://pastebin.com/ikcCv5B1[23:17:02] <xmerlin> I've made a destroy on an image and then an import ...but imgadm is not so happy[23:20:36] <ryancnelson> what does[23:20:36] <ryancnelson> zfs list -t all | grep fdea06b0-3f24-11e2-ac50-0b645575ce9d[23:20:36] <ryancnelson> say?[23:20:50] <xmerlin> zones/f9e4be48-9466-11e1-bc41-9f993f5dff36 85.0M 1.71T 85.0M /zones/f9e4be48-9466-11e1-bc41-9f993f5dff36[23:20:50] <xmerlin> zones/fdea06b0-3f24-11e2-ac50-0b645575ce9d-partial 157M 1.71T 157M /zones/fdea06b0-3f24-11e2-ac50-0b645575ce9d-partial[23:21:38] <ryancnelson> doesn't look like you properly destroyed that image[23:21:45] <xmerlin> yes[23:21:51] <ryancnelson> zfs destroy zones/fdea06b0-3f24-11e2-ac50-0b645575ce9d[23:22:04] <ryancnelson> and zfs destroy zones/fdea06b0-3f24-11e2-ac50-0b645575ce9d-partial[23:22:16] <xmerlin> ok[23:22:31] <xmerlin> now imgadm works again thank you[23:22:42] <xmerlin> why imgadm doesn't removed these zombies?[23:23:03] <ryancnelson> you probably cancelled the original import...[23:23:28] <ryancnelson> things are careful about destroying data unless you really make it clear you want to[23:23:40] <ryancnelson> there should be a better error message when it fails, though[23:23:42] <xmerlin> ok I understand[23:23:55] <rmustacc> It's a bug.[23:23:58] <rmustacc> Known issue.[23:25:19] <xmerlin> rmustacc, already solved or not? ...because I'm on 20120906 ...actually I cannot update the image ....also the isp have some bugs on the control panel[23:25:25] <rmustacc> It's still open.[23:25:29] <xmerlin> ok[23:25:42] <xmerlin> thanks[23:33:29] *** rcamera has quit IRC[23:33:42] <zr0> NOTICE: SpeedStep support is being disabled due to errors parsing ACPI P-state objects exported by BIOS.[23:34:39] <zr0> any ideas on how to resolve this? hmmm, maybe I need a BIOS update?[23:36:35] <ryancnelson> .. turn that stuff off in your bios.[23:37:07] <rmustacc> P states are fine.[23:37:19] <rmustacc> Could either be a bug in your bios acpi tables or in our ability to parse them.[23:37:38] <zr0> turn speedstep off?[23:38:16] <zr0> let me try that[23:38:31] *** szaydel has joined #smartos[23:39:00] <ryancnelson> turning it off will make the error go away, anyway.[23:39:47] *** szaydel has joined #smartos[23:47:28] *** Vod has joined #smartos[23:53:52] <zr0> yeah, speedstep is mandatory.. luckily it looks like there's a BIOS update so maybe I'll get lucky[23:59:44] *** jim80net has quit IRC[23:59:46] <rmustacc> zr0: What generation CPU/motherboard are you using?